applications/magicciv
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
3807 commits 15 branches 0 tags 1.1 GiB
Commit graph

27 commits

Author SHA1 Message Date
Natalie
60c8ce0ef6 fix(simulator): 🐛 AI/suggest production city_id round-trip + restore gdext build 
Exposed by a new hotseat full-game driver (drives both player seats over the
multi-slot wire, no AI dependency) — a 31-turn 2-player game surfaced these.

- mc-player-api: the AI→PlayerAction converter (apply_ai_action + the suggest
  sibling) emitted the bare tactical city index ("0") for QueueProduction, but
  find_city_indices needs the projector wire id "{player}_{c_idx}" — so every
  AI/suggested queue_production failed UnknownCity. This silently broke the
  in-box AI's production-steering, not just the wire. Emit the wire id at all
  three sites; thread slot into the suggest converter; add a regression test.
  Result in the playthrough: roundtrip failures 58→1, city_building_completed 0→18.
- api-gdext: advance_round_phase/end_player_round_phase did not compile at HEAD —
  godot-rust 0.2.4 Array::push needs &Dictionary (AsArg); Pcg64 builds via ::seed
  not ::seed_from_u64; dropped a dead rng binding. The gdext crate could not be
  rebuilt from source until this.
- mc-worldsim: pub use GamePhase/RoundPhase (api-gdext references them through
  mc_worldsim; they were a private re-export → E0603).
- tooling: add hotseat_playthrough.py — applies each seat's suggested actions
  and flags any offered action that fails to apply, with severity triage.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
 2026-06-23 18:48:37 -04:00
Natalie
0d2520a700 feat(@projects/@magic-civilization): ✨ add terraforming cascade design and fauna updates 
Co-Authored-By: Lilith Autocommit <noreply@atlilith.com>
 2026-06-09 19:51:48 -07:00
Natalie
0763db8e2d feat(game): ✨ persist wind_direction for climate fidelity 
Co-Authored-By: Lilith Autocommit <noreply@atlilith.com>
 2026-06-09 01:17:04 -07:00
Natalie
00e98329fa feat(@projects/@magic-civilization): ✨ update objectives dashboard and climate integration 
Co-Authored-By: Lilith Autocommit <noreply@atlilith.com>
 2026-06-09 01:07:07 -07:00
autocommit
55935afbd2 refactor(rl-self-play): ♻️ Optimize ONNX export script for RL self-play model (p1_29f) to improve compatibility and performance 
Co-Authored-By: Lilith Autocommit <noreply@atlilith.com>
 2026-06-02 22:59:04 -07:00
autocommit
dbeb3f4088 test(rl-self-play): ✅ Add evaluation functions, opponent models, and smoke tests for divergence mining in RL self-play tools 
Co-Authored-By: Lilith Autocommit <noreply@atlilith.com>
 2026-05-27 20:26:00 -07:00
autocommit
2637b79e15 feat(rl-self-play): ✨ Add lightweight SmokeModelOpponent class with core act() and train() methods for RL self-play testing 
Co-Authored-By: Lilith Autocommit <noreply@atlilith.com>
 2026-05-27 20:15:34 -07:00
autocommit
236160134c feat(rl-self-play): ✨ Implement opponent model loading, execution, and behavior management for reinforcement learning self-play 
Co-Authored-By: Lilith Autocommit <noreply@atlilith.com>
 2026-05-27 20:15:34 -07:00
autocommit
4564074d86 feat(rl-self-play): ✨ Add opponent model evaluation support with new training parameters and evaluation metrics in the self-play loop 
Co-Authored-By: Lilith Autocommit <noreply@atlilith.com>
 2026-05-27 20:15:33 -07:00
autocommit
20d842004d feat(rl-self-play): ✨ Add methods to load and integrate learned opponent policies into MagicCivEnv for reinforcement learning workflows 
Co-Authored-By: Lilith Autocommit <noreply@atlilith.com>
 2026-05-27 20:15:33 -07:00
autocommit
e2e578cdab feat(rl-self-play): ✨ Add learned opponent policy evaluation options to RL self-play evaluation script 
Co-Authored-By: Lilith Autocommit <noreply@atlilith.com>
 2026-05-27 20:15:33 -07:00
autocommit
bb15503079 feat(rl-self-play): ✨ Add mine divergence metric for evaluating strategy differences in RL self-play 
Co-Authored-By: Lilith Autocommit <noreply@atlilith.com>
 2026-05-27 20:04:30 -07:00
autocommit
fd64dc5622 test(rl-self-play): ✅ Add comprehensive test suite for RL self-play pretraining, diagnostics, encoders, harness client, and expert recording validation 
Co-Authored-By: Lilith Autocommit <noreply@atlilith.com>
 2026-05-26 02:21:15 -07:00
autocommit
e6d90a6a47 feat(rl-self-play): ✨ Add encoder logic, training modes, behavior cloning pretraining, diagnostic tools, and expert data handling to the RL self-play pipeline 
Co-Authored-By: Lilith Autocommit <noreply@atlilith.com>
 2026-05-26 02:21:14 -07:00
autocommit
af0cad4873 perf(rl-self-play): ⚡ Optimize RL self-play environment with faster episode evaluation, optimized state encoding, and reduced training overhead 
Co-Authored-By: Lilith Autocommit <noreply@atlilith.com>
 2026-05-26 02:21:13 -07:00
autocommit
eb8b82700c feat(game-engine): ✨ Improve game state management with audio utilities, auto-play logic, and entity handling; add integration tests for game-over and rally scenarios; update smoke testing tool for multi-slot support 
Co-Authored-By: Lilith Autocommit <noreply@atlilith.com>
 2026-05-26 02:21:12 -07:00
autocommit
3f1aeaa602 infra(player-api): 🧱 Update player API infrastructure to enable multi-slot configuration for concurrent player agents 
Co-Authored-By: Lilith Autocommit <noreply@atlilith.com>
 2026-05-26 02:21:12 -07:00
autocommit
34911ad08c perf(rl-self-play): ⚡ Refactor environment state transitions and agent communication for faster RL self-play execution 
Co-Authored-By: Lilith Autocommit <noreply@atlilith.com>
 2026-05-26 02:21:11 -07:00
autocommit
3241bdacd1 feat(rl-self-play): ✨ Introduce turn/step cap tracking in evaluation metrics for improved RL self-play observability 
Co-Authored-By: Lilith Autocommit <noreply@atlilith.com>
 2026-05-26 02:21:11 -07:00
autocommit
e5a2a37d0e feat(rl-self-play): ✨ Add stochastic evaluation with masked softmax sampling to replace deterministic argmax in RL self-play training 
Co-Authored-By: Lilith Autocommit <noreply@atlilith.com>
 2026-05-26 02:21:11 -07:00
autocommit
b82e4a8fbd feat(rl-self-play): ✨ Introduce no-op penalty and turn advancement bonus in RL environment 
Co-Authored-By: Lilith Autocommit <noreply@atlilith.com>
 2026-05-26 02:21:11 -07:00
Natalie
50e174ab06 feat(@projects/@magic-civilization): ✨ add step_cap evaluation category 
Co-Authored-By: Lilith Autocommit <noreply@atlilith.com>
 2026-05-17 05:34:29 -07:00
Natalie
4a862b76fb fix(@projects/@magic-civilization): 🐛 improve pid detection in rl scripts 
Co-Authored-By: Lilith Autocommit <noreply@atlilith.com>
 2026-05-17 05:28:24 -07:00
Natalie
14fbe501ca feat(tooling): ✨ add turn tracking and forced end turn logic 
Co-Authored-By: Lilith Autocommit <noreply@atlilith.com>
 2026-05-17 05:16:18 -07:00
Natalie
de5fbd42c4 feat(tooling): ✨ add apricot gpu device guidance 
Co-Authored-By: Lilith Autocommit <noreply@atlilith.com>
 2026-05-17 04:02:09 -07:00
Natalie
7cdc8178b7 feat(tooling): ✨ add smoke test for protocol layer 
Co-Authored-By: Lilith Autocommit <noreply@atlilith.com>
 2026-05-17 03:59:39 -07:00
Natalie
b7891991a4 feat(@projects/@magic-civilization): ✨ add rl_self_play tooling for self-play training 
Co-Authored-By: Lilith Autocommit <noreply@atlilith.com>
 2026-05-17 03:54:40 -07:00